1613133 - Some websites do not show video context menu for video content due to transparent overlays on the video

Status: RESOLVED DUPLICATE of bug 1611121

(Firefox :: Menus, enhancement)

Description

[Jeronimo Torti (Please send NI request to Brindusa Tot)]

In order to activate PiP mode, go to about:config

media.videocontrols.picture-in-picture.enabled needs to be set to true - To activate the feature
media.videocontrols.picture-in-picture.video-toggle.enabled needs to be set to true - To activate the toggle
media.videocontrols.picture-in-picture.video-toggle.flyout-enabled needs to be set to true - To activate the “flyout” toggle on the first <video> that users see on a page
media.videocontrols.picture-in-picture.video-toggle.flyout-wait-ms - Controls how long the “flyout” toggle waits before sliding back

• Go to following pages as examples:
  https://www.facebook.com/videos
  https://www.twitch.tv/
  https://edition.cnn.com/
• Play any available video. Notice that PiP blue toggle is available on the right side.
• While playing the video, right click over video to show context menu.
• In context menu, look for 'Picture-in-Picture' option to activate PiP mode.

Actual results:
PiP option is not available in context menu videos.

Expected:
PiP option is available in context menu videos.

Note: in https://edition.cnn.com/ PiP option from context menu is only available for one second when the video starts to load, then is another context menu the one that appears.

Comment 1

[:Gijs (he/him)]

Can you provide screenshots of the context menus you're seeing, and more details on which videos on these sites you're using to test?

On Facebook, I get a custom video context menu which we cannot change, but if I use shift+right-click to open the context menu, I get "our" context menu and that does offer a picture-in-picture option.

I think the main issue with the other two is that we don't get a video context menu at all - but it's hard to be sure because the STR and actual results do not offer enough detail (which also means I'm not sure if you get the same thing on FB as I do).

Comment 2

[Jeronimo Torti (Please send NI request to Brindusa Tot)]

Attachment: cnnf.jpg

Comment 3

[Jeronimo Torti (Please send NI request to Brindusa Tot)]

Attachment: cnnv.jpg

Comment 4

[Jeronimo Torti (Please send NI request to Brindusa Tot)]

Attachment: t.jpg

Comment 5

[Jeronimo Torti (Please send NI request to Brindusa Tot)]

Hi,
thanks for Facebook tip, I was not noticed about it.

The only pages besides FB I was able to reproduce this issue were the detailed ones:

https://www.twitch.tv/
https://edition.cnn.com/

In CNN site, PiP option from context menu is only available for one second when the video starts to load, then is another context menu the one that appears. Please, refer to screenshots 'cnnv.jpg' (when appears) and 'cnnf.jpg' (when it doesn't). Also, I attach 't.jpg' to refer how it looks context menu in Twitch site.

Regards!

Comment 6

[:Gijs (he/him)]

OK, so the screenshots match what I was seeing.

Not sure what we can do about this; it's effectively a markup/site issue, with the sites ending up with (semi-)transparent overlays over the video, which mean the content menu target is something other than the video.

I think in theory we can fix this by using a technique similar to the PiP button code for the context menu when the context menu click occurs, ie taking into account all elements at a given coordinate and checking to what degree the video or whatever is occluded. This would be a fairly significant (but positive!) change for the context menu code (which would have to keep track of several nodes, not "just" the 'target', but could also benefit other cases like bug 1611121.

Comment 7

[:Gijs (he/him)]

I'll pass the ni to Mike when he's back. :-)

Comment 8

[:Gijs (he/him)]

(In reply to :Gijs (he/him) from comment #6)

OK, so the screenshots match what I was seeing.

Not sure what we can do about this; it's effectively a markup/site issue, with the sites ending up with (semi-)transparent overlays over the video, which mean the content menu target is something other than the video.

I think in theory we can fix this by using a technique similar to the PiP button code for the context menu when the context menu click occurs, ie taking into account all elements at a given coordinate and checking to what degree the video or whatever is occluded. This would be a fairly significant (but positive!) change for the context menu code (which would have to keep track of several nodes, not "just" the 'target', but could also benefit other cases like bug 1611121.

So the question of if we can do something here and if so what, and how to prio that, over to Mike now that he's back...

Comment 9

[Mike Conley (:mconley) (:⚙️)]

I think this is a great idea for an enhancement. Were we to do this, we'd need to deal with some of the same considerations as the PiP toggle - probably the biggest one being: how do we handle transparencies over media?

I wonder if the context menu should have a submenu in that case, kind of like "This Frame", but instead "This Video", or "This Image", and then the submenu allows direct control to the underlying video or image or what-have-you.

I think that'd be pretty neat, and a great way of empowering our users over web content.

Would be worth asking UX what they'd want to do in the context menu for these cases, and maybe this could be a nice small-medium sized project for somebody.

Comment 10

[Mike Conley (:mconley) (:⚙️)]

This is basically a dupe of bug 1611121 now. :)